Deciphering the Allosterically Driven Conformational Ensemble in

Subscriber access provided by KEAN UNIV

Article

Deciphering the Allosterically-driven Conformational Ensemble in Tryptophan Synthase Evolution Miguel A. Maria-Solano, Javier Iglesias-Fernandez, and Sílvia Osuna J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.9b03646 • Publication Date (Web): 29 Jul 2019 Downloaded from pubs.acs.org on July 29, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Deciphering the Allosterically-driven Conformational Ensemble in Tryptophan Synthase Evolution Miguel A. Maria-Solano,[a] Javier Iglesias-Fernández,*[a] and Sílvia Osuna*[a,b] [a] CompBioLab group, Institut de Química Computacional i Catàlisi (IQCC) and Departament de Química, Universitat de Girona, Girona, Spain [b] ICREA, Pg. Lluís Companys 23, 08010 Barcelona, Spain

ABSTRACT: Multimeric enzyme complexes are ubiquitous in Nature and catalyze a broad range of useful biological

transformations. They are often characterized by a tight allosteric coupling between subunits, making them highly inefficient when isolated. A good example is Tryptophan synthase (TrpS), an allosteric heterodimeric enzyme in the form of an αββα complex that catalyzes the biosynthesis of L-tryptophan. In this study, we decipher the allosteric regulation existing in TrpS from Pyrococcus furiosus (PfTrpS), and how the allosteric conformational ensemble is recovered in laboratory-evolved stand-alone β-subunit variants. We find that recovering the conformational ensemble of a subdomain of TrpS affecting the relative stabilities open, partially-closed and closed conformations is a prerequisite for enhancing the catalytic efficiency of the β-subunit in the absence of its binding partner. The distal mutations resuscitate the allosterically-driven conformational regulation and alter the populations and rates of exchange between these multiple conformational states, which are essential for the multistep reaction pathway of the enzyme. Interestingly, these distal mutations can be a priori predicted by careful analysis of the conformational ensemble of the TrpS enzyme through computational methods. Our study provides the enzyme design field with a rational approach for evolving allosteric enzymes towards improved stand-alone function for biosynthetic applications.

INTRODUCTION Allostery is a central biological process in which two distinct sites within a biomolecule are functionally connected. Allosteric effects play a key role in protein regulation and cell signaling, and their functional significance has fostered many studies for unveiling the underlying forces that drive allostery. 1-3 In enzymatic mechanisms, allosteric interactions often promote enzyme-substrate binding and product release, and directly affect catalytic turnover.4-6 Some studies suggest that allostery is an intrinsic characteristic of enzymes,7 given the fact that distal active site mutations often confer improved catalytic properties.8-9,10-11 The essential role played by remote mutations in tuning enzyme activity also indicates that allostery could be exploited for the engineering of new enzyme variants.12

Allosteric regulation present in multimeric enzyme complexes makes the isolated subunits, i.e. in the absence of their protein partner, highly inefficient. This is indeed the case for Tryptophan synthase (TrpS; EC 4.2.1.20). TrpS is a heterodimeric enzyme complex composed of α-subunits (TrpA) and β-subunits (TrpB) in an αββα arrangement that presents an intricate allosteric communication network between TrpA and TrpB.13-15 TrpA catalyzes the retro-aldol cleavage of indole glycerol phosphate (IGP) producing glyceraldehyde 3-phosphate (G3P) and indole; the latter is able to diffuse through an internal TrpA-TrpB tunnel to reach the TrpB subunit (see Figure 1).

ACS Paragon Plus Environment


β-subunit (TrpB)

O

NH 3

O OH

H

N H N

L-Tryptophan

O

H

O

P NH

Indole

3

OH

HO

E(Q2), which after proton extraction (to recover indole’s aromatic character) generates E(Q3) (not shown in Figure 2a). At this point, protonation at Cα of Q3 by K82 forms the E(Aex2) intermediate, which undergoes a second transamination reaction to finally release the Ltryptophan E(Trp) product and restore the enzyme resting state (see Figure 2).15 Previous studies along the catalytic mechanism identified different open and closed conformations of the enzyme in both subunits, which were based on static Xray structures. These open-to-closed transitions can be defined by the TrpA loop (residues 163-176) that gets ordered and the slow motion of the rigid COMM domain in TrpB (residues 97-184, see Figure 2 and Table S1). Both the TrpA loop and the TrpB COMM domain are part of the active site cavity of each subunit and modulate solvent exposure to prevent substrate loss to the media. Besides, the COMM domain contains the -helix H6 (residues 174-164) that is directly involved in noncovalent interactions with the indole moiety of the TrpB reactant intermediates. Moreover, E104 located in the COMM domain has been reported to play a role in the stabilization of charge redistribution that takes place during the nucleophilic attack of indole in the A-A intermediate (see Figure 2).15

O

P

OH

L-Serine

N H

Indole-3-glycerol phosphate

α-subunit (TrpA)

Figure 1. Overview of the Tryptophan synthase (TrpS) mechanism. The enzyme is a heterodimeric complex formed by two subunits: TrpA (shown in teal) and TrpB (in purple). The TrpB resting state is characterized by a pyridoxal phosphate (PLP)-cofactor covalently linked to the K82 active site residue, forming a Schiff base intermediate (E(Ain)). After transamination with L-serine E(Ser), an external aldimine intermediate E(Aex1) is formed. This intermediate undergoes deprotonation at Cα, assisted by K82, which is followed by a rapid elimination of the Aex1 hydroxyl group to form an electrophilic amino acrylate intermediate E(A-A). In the dimeric complex, indole formed in TrpA reaches the TrpB active site and reacts with E(A-A) to form a quinonoid intermediate

(a)

(b)

O HO H N

H

NH 2

K 82

L-Serine

H

H O

O O

K 82 N

P

K 82

H O

P

N

H

H O

O

N OH

P N

PC

N

C

P12L

E(A-A)-C

E(Aex1)-PC

E(L-Ser)-O

O

O

O

P N

E(Ain)-O

HO H NH 2 N

L-Tryptophan

H N O H 2N

C

H O

H O

N

O H

N

PC

E(L-Trp)-PC

N

N

O

N

H

N

OH

I68V OH

P

E(Aex2)-C

O

O H

H OH

P

P

T321A

H

K 82 K 82

PLP

H6 O

O

N

T292S

O

H N

O

H

F274S

E104

O H N

PC C

E17G

E104 O

NH 2

COMM DOMAIN

O

H 2O

Indole

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Glyceraldehyde phosphate

Page 2 of 10

P N

E(Q2)-C

C

N

C

E(A-A,Indole)-C

Figure 2. (a) Tryptophan synthase reaction mechanism of TrpB subunit showing the conformational states of the COMM domain according to available X-ray data at each reaction intermediates. The degree of closure of the COMM domain is represented by colored labels in blue (open, O), teal (partially closed, PC), and green (closed, C). (b) Overlay of representative X-ray structures showing the COMM domain(97-184 residues) in O, PC and C states. The pyridoxal phosphate (PLP) cofactor is shown in purple, the PfTrpB0B2 stand-alone DE mutation positions are marked with purple spheres, and the COMM domain -helix H6 (residues 174-164) is highlighted.


Page 3 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Tryptophan synthase has found applications in many fields of synthetic chemistry, in particular, for the production of noncanonical amino acids (NCAAs).16-18 The use of TrpS for industrial purposes is hampered by its multimeric structure and the low activity of TrpB as stand-alone enzyme. Detailed insights were obtained in studies by Prof. Arnold and co-workers, who applied Directed Evolution (DE) to a thermophilic TrpB from Pyrococcus furiousus (PfTrpB). They optimized the enzyme for stand-alone function,19 and later on for the production of a variety of Trp derivatives.19-25 The most efficient stand-alone catalyst was achieved by introducing up to six activating mutations, which were located far away from active site positions (see Figure 2b). Note that P12L, E17G and F274S are located close to the TrpA-TrpB protein interface. Analysis of spectroscopic data suggested that PfTrpB stand-alone variants and the PfTrpS complex were better in stabilizing closed conformations of the COMM domain upon substrate binding than isolated PfTrpB.24 However, despite showing drastic differences in catalytic efficiency, X-ray data failed to find a connection between COMM domain closure and stabilization of the enzyme. In particular, the COMM domain structure is almost identical among different organisms (e.g. Salmonella typhimurium and Pyrococcus furiosus), isolated PfTrpB enzyme and PfTrpB stand-alone variants, although all of them diverge in functionality. These observations suggest that the origin behind their different catalytic efficiencies could be attributed to alterations in the enzyme conformational ensemble induced by distal active site mutations. Such effects have not been explored yet, although they are crucial to understand how the stand-alone functionality was achieved. In this work, we elucidate how the different reaction intermediates and distal mutations introduced in labor-

atory-evolution alter the allosterically-driven conformational ensemble of PfTrpS. Surprisingly, the introduced distal mutations increase the conformational heterogeneity of the COMM domain; hence, the PfTrpB enzyme has the ability to access the different COMM domain conformations, which are essential for efficient catalysis in the absence of its binding partner. Through careful inspection of the conformational ensemble of PfTrpS with our recently developed SPM tool12 we were able to identify the most important positions to recover the allosterically-driven conformational ensemble, which coincide with the mutations introduced in laboratoryevolution. Our study shows clearly that stand-alone versions of allosterically-regulated enzymes can be rationally designed by targeting the recovery of the allosterically-driven conformational ensemble.

RESULTS AND DISCUSION Available structural data show that the TrpB COMM domain is able to explore open (O), partially closed (PC), and closed (C) conformations along the multi-step TrpB catalytic pathway, due to the allosteric regulation exerted by TrpA (see Figure 2). Considering only the TrpB subunit, X-ray studies revealed that its resting state (i.e. E(Ain)) is characterized by O COMM domain conformations (1V8Z)26, which are shifted towards PC states at the external aldimine intermediate E(Aex1) (5DW0)19. All subsequent reaction intermediates (i.e. from the electrophilic amino acrylate E(A-A) to E(Aex2)) were crystallized in C states (4HN427 and 3CEP28). A recent Xray structure (5DW3)19 indicated that the PC conformation is recovered once Trp is formed at E(Trp), preparing the enzyme for product release and the next turnover (see Figure 1-2 and Table S1 for more structural data).

PfTrpS αβ complex

(a)

Aex1

Ain

O

(b)

Q2

PfTrpS αβ complex (Ain)-O

3D

O

(c)

PC

O

(Aex1)-PC (Q2)-C

PC

O

H6

PC C

I165

PfTrpS(Q2)-C E104 αβ complex L161 H110

3.7±0.3

G184 3.6±0.2

2D

C PC

O

K82

H6

O PC

O

PC

Q2

Y301 S185

Q2

Figure 3. (a) Free energy landscape (FEL) associated to the COMM domain Open-to-Closed (O-to-C) conformational exchange of the PfTrpS complex enzyme at Ain, Aex1 and Q2 reaction intermediates. (b) Overlay of the PfTrpS metastable conformations of the open (O) state at Ain intermediate, partially closed (PC) at Aex1 and closed (C) at Q2, respectively showing the entire O-to-C sampled transition. (c) Detailed active site view of the PfTrpS metastable conformation of the C state at Q2 intermediate (shown in purple). Active site residues are shown in gray, except for those included in the COMM


Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 10

domain (shown in blue), and the catalytic K82 proton transfer residue (green). The catalytic distances (in Å) between charge-charge stabilization E104-Q2 and proton transfer K82-Q2 are also represented.

Allosteric transitions, such as the TrpA-triggered O-toC exchange of the COMM domain in TrpB, are relatively slow domain motions that take place on time scales larger than our currently accessible simulation times.29 Indeed, initial 500 ns standard MD simulations of the PfTrpS in the αβ complex, the isolated PfTrpB wild-type and the stand-alone PfTrpB0B2 enzyme variant in multiple reaction intermediates (Ain, Aex1 and A-A) failed to sample the entire allosteric transition, with no clear RMSD differences observed between the studied systems (Figure S2). To overcome this limitation, we employed enhanced sampling techniques. In particular, we applied the metadynamics approach30-31 to reconstruct the Free Energy Landscape (FEL) associated with the COMM domain O-to-C transition of the PfTrpS αβ complex, isolated PfTrpB wild-type, and evolved stand-alone PfTrpB0B2 variant (see details in SI). Several intermediates along the catalytic cycle were modeled. In particular, we selected E(Ain), E(Aex1), E(A-A) and E(Q2) to evaluate the O-PC-C conformational exchange of the COMM domain found in X-ray data, but also to reproduce the multi-step mechanism under study (see Figure 2).

Population shift towards closed conformations along the allosteric TrpS catalytic pathway. To elucidate the allosterically-driven conformational ensemble of PfTrpS αβ complex we reconstructed the FEL associated to the conformational dynamics of the COMM domain for each reaction intermediate (see Figure 3a). As expected from X-ray data, in the resting state of the enzyme, PfTrpS-Ain, the O conformational state is highly favored, in agreement with its functional role in Ser binding. However, less stable PC states (ca. 2 kcal/mol higher in energy) are also visited with an associated O-to-PC transition energy barrier of only ca. 3 kcal/mol. As the PfTrpS enzymatic reaction progresses, a population shift occurs towards the stabilization of PC states (see Figure 3a). After the reaction with serine in the external aldimine Aex1 intermediate, the open O state is destabilized by ca. 2 kcal/mol with respect to the PC state, which becomes the most stable

conformation. In contrast to Ain and Aex1, the quinonoid Q2 intermediate generated after indole coupling samples all possible conformations of the COMM domain: O and PC states are almost equally stabilized, while the C state is ca. 5 kcal/mol higher in energy. The associated PC-to-C barrier is ca. 6 kcal/mol. This suggests that the adoption of the fully closed COMM domain conformation is the limiting factor, in agreement with the spectroscopic data for PfTrpS.19, 24 Such closed active states form an optimized network of hydrophobic interactions between the enzyme and the indole moiety (see Figure 3c). Several side chains residues, including Y301, S185, and COMM domain G184, H110, L161 and I165 define this network. Comparison of PfTrpS(Q2)-O and -C metastable structures shows that the helix H6 closure is needed for forming CH··CH and CH··π interactions between L161 and I165 with the indole moiety and also a hydrogen bond with the E104 residue (Figure 3b-c). The C state of PfTrpS(Q2) shows a highly pre-organized active site with E104 and the proton acceptor K82 properly positioned for catalysis together with the indole moiety establishing many non-covalent interactions with the active site pocket (see Figure 3c and S8a-b). The high stability of O states at the Q2 intermediate suggests that the COMM domain of PfTrpS(Q2) is already prepared for product release and recovery of the native state of the enzyme for the next cycle. Altogether, these findings highlight the crucial role of the allosterically-driven conformational ensemble of the COMM domain of PfTrpS for efficiently optimizing the multiple steps along its catalytic cycle.


Page 5 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(a)

(b)

PfTrpB isolated Ain

Aex1

(c)

Q2

3D C O

PfTrpB(Q2)-C isolated PfTrpS(Q2)-C αβ complex

H6

PfTrpB(Q2)-C E104 isolated

PC

I165

3.9±0.3

2D

3.9±0.2

Y301

H6 O

L161 H110

PC

G184 K82

C

S185 Q2

Q2

Figure 4. (a) Free energy landscape (FEL) associated to the COMM domain Open-to-Closed (O-to-C) conformational exchange of the PfTrpB isolated enzyme at Ain, Aex1 and Q2 reaction intermediates. (b) Overlay of the metastable conformations of the closed (C) states at Q2 intermediate for PfTrpB (in orange) and PfTrpS (blue). (c) Detailed active site view of the PfTrpB metastable conformation of the C state at Q2 intermediate (shown in purple). Active site residues are shown in gray, except for those included in the COMM domain (shown in orange), and the catalytic K82 proton transfer residue (green). The catalytic distances (in Å) between charge-charge stabilization E104-Q2 and proton transfer K82-Q2 are also represented.

Isolated PfTrpB displays restricted COMM domain heterogeneity and unproductive closure. Experimental data showed that in the absence of the allosteric partner PfTrpA, PfTrpB activity decreases 3-fold (kcat of 0.31 and 1.0 s-1 for isolated PfTrpB and PfTrpS, respectively).19 Our reconstructed FELs corresponding to isolated PfTrpB display some similarities to the PfTrpS system (see Figures 3a and 4a). Ser binding at the Aex1 intermediate shifts the conformational ensemble from O towards PC states. Similarly, a population shift towards C states at the Q2 intermediate is observed. Contrary to the situation for the PfTrpS complex, a single energy minimum is found at the Aex1 and Q2 intermediates of PfTrpB. In fact, the COMM domain is not able to escape from O states at Ain, PC at Aex1, and C at Q2 intermediates as the other states are inaccessible. Therefore, PfTrpB in the absence of PfTrpA allosteric regulation has a very limited conformational heterogeneity of the COMM domain, which hampers the multi-step reaction pathway. It is also worth mentioning that the stable C states at the PfTrpB-Q2 intermediate are highly deviated from the reference O-to-C conformational path (i.e. RMSD larger than 1.5 Å, see Figure 4a). A detailed structural analysis of the isolated wild-type PfTrpB(Q2) compared to the PfTrpS(Q2) complex in C states, indicates that the isolated PfTrpB enzyme cannot efficiently sample catalytically competent C states; this is in particular true for the key COMM H6 closure (Figure 4b-c and S8c). Furthermore, the proton transfer catalytic distance K82-Q2 is also longer than in PfTrpS (3.9±0.3 Å vs.

3.6±0.3 Å). Our simulations have therefore shown that in the absence of its PfTrpA allosteric partner, the PfTrpB COMM domain displays a restricted conformational landscape, which lacks the ability to easily access O, PC, and C states existing in the allosterically-driven conformational ensemble of PfTrpS.

Activating distal mutations for stand-alone function recover COMM domain heterogeneity. PfTrpB was evolved for stand-alone function generating a new variant PfTrpB0B2, which displays a considerably improved catalytic constant with respect to both the isolated wild-type PfTrpB and the PfTrpS complex (kcat of 2.9 s-1, 0.31 s-1, 1.0 s-1 for PfTrpB0B2, PfTrpB, PfTrpS, respectively). It is also worth mentioning that the activity of the evolved PfTrpB0B2 decays dramatically in the presence of PfTrpA (kcat of 0.04 s-1).19 Intrigued by the restricted conformational dynamics of the COMM domain as found in isolated PfTrpB, we decided to explore whether distal mutations introduced in laboratory evolution were able to recover the allosterically-driven conformational ensemble of PfTrpS. By comparing the reconstructed FELs for stand-alone PfTrpB0B2 and PfTrpS complex along the different reaction intermediates (see Figure 3a and 5a), it becomes evident that the PfTrpB0B2 variant recovers the conformational heterogeneity of the COMM domain, characteristic of the allosterically regulated dimeric enzyme. However, interesting differences between both systems are found to be crucial for rationalizing their catalytic activities.



(a)

(b)

PfTrpB0B2 stand-alone Ain

PC

O

PC

(c)

Q2

Aex1

3D

O

Page 6 of 10

PfTrpB0B2(Q2)-C stand-alone PfTrpS(Q2)-C αβ complex

O C

2D

PfTrpB0B2(Q2)-C stand-alone H6 E104 L161 H110 3.9±0.2 I165 G184 3.6±0.3

H6 O

O

PC

O

PC

K82

C

Q2

Y301

S185

Q2

Figure 5. (a) Free energy landscape (FEL) associated to the COMM domain Open-to-Closed (O-to-C) conformational exchange of the PfTrpB0B2 enzyme at Ain, Aex1 and Q2 reaction intermediates. (b) Overlay of the metastable conformations of the closed (C) states at Q2 intermediate for PfTrpB0B2 (in pink) and PfTrpS (blue). (c) Detailed active site view of the PfTrpB0B2 metastable conformation of the C state at Q2 intermediate (shown in purple). Active site residues are shown in gray, except for those included in the COMM domain (shown in violet), and the catalytic K82 proton transfer residue (green). The catalytic distances (in Å) between charge-charge stabilization E104-Q2 and proton transfer K82-Q2 are also represented.

PfTrpB0B2

In the resting state (Ain), has only the O state accessible. However, as the reaction progresses, a population shift towards PC and C states occurs, as also observed in the dimeric PfTrpS complex. At Aex1, O and PC states have comparable relative stabilities and are separated by a small energy barrier of ca. 1 kcal/mol, which allows a fast O-to-PC conformational exchange. Similar to the PfTrpS system, at the Q2 state the allosterically-driven conformational ensemble containing O, PC, and C states is recovered. Nevertheless, a substantially lower barrier is observed for the O-to-PC-to-C transition of ca. 2 kcal/mol as compared to PfTrpS complex. This rather small energy barrier allows PfTrpB0B2 to easily adopt the catalytically competent C conformation from PC and O states. This high stability of the catalytically relevant C state contrasts with the PfTrpS system where the closed state is ca. 5 kcal/mol higher in energy. Such a difference in the stability of the C state explains the improved catalytic efficiency of the evolved stand-alone variant. The C state of stand-alone PfTrpB0B2(Q2) has an almost identical degree of closure of the COMM domain as the PfTrpS catalytically competent conformation, and a similar catalytic K82-Q2 proton transfer distance (see Figure 5b-c and S9). This indicates that the C state of the stand-alone PfTrpB0B2(Q2) variant is properly preorganized for the reaction. A remarkable difference between the dimeric PfTrpS(Q2) complex and stand-alone PfTrpB0B2(Q2) is found at the PC state, which in the case of the evolved variant is highly digressed from the original path (i.e. RMSD > 1.5 Å, see Figure 5a); therefore, we denote this as the novel PC state. The large deviation arises from an unexpected large-scale conformational change of 14 Å

that positions R159, adjacent to the H6 helix of the COMM domain, towards the active site (see Figure 6a). This novel PC conformation has not been previously observed by means of X-ray crystallography. Interestingly, R159 takes over the position previously occupied by L161 at H6 of PfTrpB0B2(Q2), establishing a cation-π interaction with the indole moiety (Figure 6b). We hypothesize that this novel conformation of R159 may play a role in the catalytic cycle, most probably in properly positioning serine and/or indole for the reaction (see SI for a detailed discussion and Figure S10-11).

(a)

(b)

P f TrpB0B2(Q2)-C and -PC stand-alone

P f TrpB0B2(Q2)-PC stand-alone

H 6

H 6

R 1 5 9

Q 2

Figure 6. (a) Overlay of the PfTrpB0B2 metastable conformations of the closed (C, in purple) and novel partially closed (PC, in green) states at the Q2 reaction intermediate. The novel PC state revealed by MD simulations presents R159 from the COMM domain located in the active site close to the position previously occupied by L161 in the C state. (b) Representation of the noncovalent interactions (computed with NCIplot32 at the novel PC state of PfTrpB0B2, highlighting (green surfaces) cation-π interaction between R159 and the indole moiety of Q2.


Page 7 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Our findings indicate that stand-alone PfTrpB0B2 recovers the COMM domain heterogeneity, thus making O and active C states accessible again. The great stabilization of active C states together with the novel PC conformation displayed by the PfTrpB0B2 enzyme variant unravel the increase in catalytic efficiency with respect to the allosterically regulated PfTrpS complex. Another relevant aspect is the experimentally observed inactivation of the evolved PfTrpB0B2 in the presence of PfTrpA. To study this inactivation, we reconstructed the FEL for the PfTrpA-PfTrpB0B2(Q2) complex. Surprisingly, our simulations show that the presence of PfTrpA does not restrict the PfTrpB0B2 COMM domain heterogeneity, as it is also able to sample the O-to-C exchange. However, the formation of the dimeric complex with PfTrpA induces a population shift towards unproductive closed states (i.e. highly deviated from the reference path), similar to those observed in the isolated PfTrpB system (see Figure S12). Thus, PfTrpA truncates the efficient conformational ensemble of the stand-alone PfTrpB0B2 yielding non-productive closed conformational states of the COMM domain.

COMM domain heterogeneity as an essential factor in indole active-site accessibility. Available X-ray structures after E(Aex1) formation display C conformations of the COMM domain. However, our analysis of the allosterically-driven conformational ensemble of PfTrpS complex and that of stand-alone PfTrpB0B2 provided evidence for a high flexibility of the COMM domain and the ability to visit O, PC, and the catalytically relevant C states. The question that remains is: what is the specific role of O conformational states of the COMM domain after Ser binding? One possibility would be to assist in either indole binding or Trp release after the reaction. Experimentally, the Michaelis constant for indole binding in the stand-alone PfTrpB0B2 enzyme variant (8.7 μM) was improved with respect to isolated PfTrpB (77 μM), but also compared to the enzyme complex PfTrpS (20 μM). 19 To elucidate the changes in indole binding and the role played by the COMM domain O states, we reconstructed the FELs at the electrophilic amino acrylate E(A-A) intermediate (Figure 7a-b and S13), and analyzed the available indole substrate access tunnels with the CAVER software.33 At this A-A intermediate, PfTrpS complex and stand-alone PfTrpB0B2 can easily access both O and C states, which are separated by relatively small energy barriers. The analysis of indole access tunnels in both O and C states reveals two different entry pathways (Figure 7c-d and S13c): the previouslydescribed internal tunnel (IT) that connects TrpA and TrpB subunits in PfTrpS complex (shown in blue in Figure 7c-d and S13c), and a secondary tunnel (ST) connecting the active site with a novel entry path not de-

scribed before (shown in green). C states of the COMM domain yield a narrow bottleneck tunnel radius hampering indole diffusion outside the active site, thus capturing it for efficient catalysis (see Figure 7-d). Therefore, the differences in indole binding should be related to O COMM domain states. Interestingly, the isolated wild-type PfTrpB is not able to sample the O state, which results in indole access through PC conformations that have a much narrower bottleneck radius (see Figure S13c). This leads to less favorable KM values for PfTrpB, as observed experimentally. At the O state of PfTrpS complex, indole diffusion occurs along the internal TrpA-TrpB tunnel, suggesting that the secondary tunnel (green in Figure 7c-d and S13C) may play a role in Ser binding and/or Trp release. For the stand-alone PfTrpB0B2 variant, both tunnels show a large bottleneck radius, thus no tunnel preference for indole entrance to the active site is found (see Figure 7c). Altogether, these calculations indicate that the recovery of the allosterically-regulated COMM conformational ensemble of PfTrpS, especially O state accessibility, is also key for indole binding.

(a)

(b)

P f TrpB0B2(A-A) stand-alone

P f TrpB0B2(A-A)-O and -C

O C

H 6

(c)

P f TrpB0B2(A-A)-O

A A

(d) P f TrpB0B2(A-A)-C

A A

A A

H 6 H 6

S T IT ( 1 . 6 5 ± 0.24) ( 1 . 7 8 ± 0.25)

IT ( 1 . 5 0 ± 0.20)

S T ( 1 . 1 5 ± 0.15)

Figure 7. (a) Free energy landscape (FEL) associated to the COMM domain Open-to-Closed (O-to-C) conformational exchange of stand-alone PfTrpB0B2 enzyme at A-A reaction intermediate. (b) Overlay of meta-stable conformations of the O (dark purple) and C state (light purple) at A-A intermediate of PfTrpB0B2. (c) and (d) Internal (IT, in blue) and secondary (ST, in green) tunnels of PfTrpB0B2 at the O and C states at A-A reaction intermediate computed with CAVER 3.0.33 The averaged bottleneck radii (in Å) are also shown.

Distal mutations for stand-alone function can be predicted computationally. Our group has recently shown that distal mutations found by DE in the case of multi-step retro-aldolase enzymes can be identi-



fied with residue-by-residue correlation and proximity analysis tools.12 Intrigued by the possibility of predicting distal positions for stand-alone function, we applied our Shortest Path Map (SPM) method.12 This computational tool identifies those pairs of residues that have a higher contribution to the conformational dynamics of the enzyme (see Figure 8 and computational details). We focused our analysis on the PfTrpS(Q2) metadynamics trajectory because of the complete O-to-C conformational exchange sampled in it (see Figure S14 for SPM analysis at other reaction intermediates). PfTrpB0B2 presents 6 mutations: P12L, E17G, I68V, T292S, F274S and T321A, from which 2 were directly predicted by the SPM tool, and 3 were directly interacting with a SPM position. The specific effect of each isolated mutation on the enzyme activity is not known, except for two of them: T292S and P12L. The most beneficial mutation T292S (3-fold increase in kcat with respect to PfTrpB)19 was previously suggested to modulate COMM domain closure based on the T292-D300 interaction observed in X-ray data.15 In our metadynamics simulations, the hydrogen bond between T292 and D300 is maintained 86% of the time (see Figure 8). Although position 292 is not directly included in our computed SPM path, position 300 is predicted as key for the COMM domain conformational dynamics. This indicates that by altering position D300 interactions (for instance, the T292-D300 interaction) the COMM domain closure can be modulated. Interestingly, P12L distal mutation P12L, which was found to have a slight impact on the kcat of the enzyme, is directly identified with high contributions in our SPM analysis (see orange spheres in Figure 8). Simi-

Page 8 of 10

larly, the distal site E17G is also predicted by SPM, suggesting a role on COMM domain conformational heterogeneity. DE positions F274, and I68, although not strictly included in the SPM path, make direct and stable non-covalent interactions with already SPM predicted positions (see blue and pink spheres in Figure 8). For instance, F274 highly forms CH··π and π ··π interactions with the SPM residues F281 (maintained 99% of the simulation time), Y301 (37%), and P302 (33%), as well as a hydrogen bond with M277 (28%). Similarly, DE position I68 makes CH··π interactions with Y69 (74%), included in SPM. The only DE position that has a minor role in the COMM domain conformational dynamics and makes negligible interactions with SPM residues is T321. Our new proposed methodology makes use of metadynamics simulations to enforce the sampling of the allosterically-regulated O-to-C transition, and identifies which residues present a higher contribution to the O-to-C COMM domain conformational exchange through inter-residue correlation calculations. With this new computational approach, distal positions involved in the allosteric transition can be identified, thus providing a set of key positions for the generation of smart libraries for stand-alone function. This new proposed protocol can be applied to any allosterically-regulated system of interest. This study also provides further evidence for the key role played by the enzyme conformational dynamics in the evolution of enhanced catalytic activities, especially in challenging multi-step mechanisms such as the one catalyzed by TrpS.34

Figure 8. Identification of the amino-acids that contribute to the Open-to-Closed (O-to-C) conformational exchange in PfTrpS at (Q2) intermediate through Shortest Path Map (SPM) analysis.12 The size of the spheres and black edges are indicative of the importance of the position for the PfTrpS conformational dynamics. Positions mutated via DE are marked in orange (if they are included in the SPM), or in pink (if they H-Bond M 2 7 7 (28%) directly interact with SPM residues). SPM residues that E 1 7 G F 2 7 4 CH··π 3 0 1 (99%) Y interact with the DE positions Y 6 9 are marked with blue spheres. π··π (57 %) CH··π For each mutation that interacts K 8 2 F 2 8 1 P (74 %) 3 0 2 with SPM residues, a zoom is P 1 2 L provided to show the type of I 6 8 M 2 7 7 M 2 7 7 non-covalent interaction and the Q 2 F 2 7 4 Y 3 0 1 percentage of interacion time P 3 0 2 Y 6 9 π··π 2 7 4 S (37%) Y during the simulation. 3 0 1F CH··π

F 2 8 1

(33 %)

F 2 8 1

D 3 0 0E 2 5 1

I 6 8 V

T 2 9 2 S

P 3 0 2 D 3 0 0

D 3 0 0

CONCLUSIONS T 3 2 1 A

E 2 5 1

Recovering the allostericallydriven conformational ensemble existing in multimeric en-

H-Bond (42 %)

T 3 2 1

H-Bond (44%)

H-Bond (8%)

T 2 9 2 T 2 9 2


Page 9 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


zymes such as TrpS for stand-alone function is strikingly similar to the dramatic effect induced by distal mutations on the catalytic efficiency of some enzymes. Only those ensembles of conformations that are preactivated for catalysis are selected and stabilized along the evolutionary process. Understanding the differences between both processes is highly appealing for the rational design of enzymes. The present study demonstrates that fine-tuned control of the allostericallydriven conformational ensemble of PfTrpS plays a key role along its catalytic cycle. By altering the relative stabilities of open, partially-closed and closed conformational states of the COMM domain, each reaction step along the catalytic pathway can be efficiently optimized. Our free energy calculations on the conformational exchange of the COMM domain indicate that the rate for the open-to-closed conformational transition is relatively fast (in the nanosecond to microsecond timescale) in comparison with the reaction steps and turnover time scale (millisecond to second). However, such transitions are essential for pre-organizing the active site pocket to accommodate the different substrates, and efficiently catalyzing Ser and indole coupling for Trp production. Distal mutations, introduced experimentally for converting PfTrpB into an efficient stand-alone variant, recover the allosterically-regulated conformational ensemble of PfTrpS. This enables access to open, partially closed and closed states of the COMM domain. In the absence of such mutations, the isolated PfTrpB lacks the COMM domain conformational heterogeneity, which is required for the challenging multi-step catalytic pathway. By careful analysis of the open-to-closed conformational exchange of the COMM domain, and the residues that contribute more to the exchange, distal mutations introduced via Directed Evolution can be predicted with our recently developed SPM tool. This study shows that by evaluating the native allosterically-regulated conformational ensemble, and the residues that have a higher contribution to the allosteric conformational transition, proficient stand-alone enzyme variants could be rationally designed. The hypothesis that many enzymes are intrinsically regulated allosterically7 is inspiring, as it also suggests that our novel computational approach proposed here might be of general use in the computational enzyme design field. ASSOCIATED CONTENT

Supporting information Computational methods as well as additional tables and figures. This material is available free of charge via the Internet at http://pubs.acs.org

Corresponding Author *[email protected] *[email protected]

ORCID Sílvia Osuna: 0000-0001-9458-1114 Javier Iglesias-Fernandez: 0000-0001-7773-2948 Miguel A. Maria-Solano: 0000-0002-7837-0429

ACKNOWLEDGMENT We thank the Generalitat de Catalunya for the emerging group CompBioLab (2017 SGR-1707) and Spanish MINECO for project PGC2018-102192-B-I00. We are grateful for the computer resources, technical expertise, and assistance provided by the Barcelona Supercomputing Center - Centro Nacional de Supercomputación. M. A. M. S. was supported by the Spanish MINECO for a PhD fellowship (BES2015-074964), J. I. F. was supported by the European Community for Marie Curie fellowship (H2020-MSCA-IF2016-753045). S.O. is grateful to the funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (ERC-2015StG-679001). We thank Prof. Swart, Dr. Feixas and Dr. Kress for the helpful discussions and comments on the manuscript.

REFERENCES (1) Motlagh, H. N.; Wrabl, J. O.; Li, J.; Hilser, V. J. Nature 2014, 508 (7496), 331-9. (2) Nussinov, R.; Tsai, C. J.; Ma, B. Annu. Rev. Biophys. 2013, 42, 169-89. (3) Freiburger, L. A.; Baettig, O. M.; Sprules, T.; Berghuis, A. M.; Auclair, K.; Mittermaier, A. K. Nat. Struct. Mol. Biol. 2011, 18 (3), 288-94. (4) Lisi, G. P.; Loria, J. P. Curr. Opin. Struct. Biol. 2017, 47, 123-130. (5) Tzeng, S. R.; Kalodimos, C. G. Nature 2012, 488 (7410), 236-40. (6) Nussinov, R. Chem. Rev. 2016, 116 (11), 6263-6. (7) Gunasekaran, K.; Ma, B.; Nussinov, R. Proteins: Struct., Funct., and Bioinf. 2004, 57 (3), 433-443. (8) Obexer, R.; Godina, A.; Garrabou, X.; Mittl, P. R. E.; Baker, D.; Griffiths, A. D.; Hilvert, D. Nat. Chem. 2017, 9 (1), 50-56. (9) Currin, A.; Swainston, N.; Day, P. J.; Kell, D. B. Chem. Soc. Rev. 2015, 44 (5), 1172-1239. (10) Morley, K. L.; Kazlauskas, R. J. Trends Biotechnol. 23 (5), 2005, 231-237.

AUTHOR INFORMATION ACS Paragon Plus Environment


(11) Jiménez-Osés, G.; Osuna, S.; Gao, X.; Sawaya, M. R.; Gilson, L.; Collier, S. J.; Huisman, G. W.; Yeates, T. O.; Tang, Y.; Houk, K. N. Nat. Chem. Biol. 2014, 10 (6), 431-436. (12) Romero-Rivera, A.; Garcia-Borràs, M.; Osuna, S. ACS Catal. 2017, 7 (12), 8524-8532. (13) Hyde, C. C.; Ahmed, S. A.; Padlan, E. A.; Miles, E. W.; Davies, D. R. J. Biol. Chem. 1988, 263 (33), 1785771. (14) Lee, S. J.; Ogasahara, K.; Ma, J. C.; Nishio, K.; Ishida, M.; Yamagata, Y.; Tsukihara, T.; Yutani, K. Biochemistry 2005, 44 (34), 11417-11427. (15) Dunn, M. F. Arch. Biochem. Biophys. 2012, 519 (2), 154-166. (16) Barry, S. M.; Kers, J. A.; Johnson, E. G.; Song, L. J.; Aston, P. R.; Patel, B.; Krasnoff, S. B.; Crane, B. R.; Gibson, D. M.; Loria, R.; Challis, G. L. Nat. Chem. Biol. 2012, 8 (10), 814-816. (17) Kieffer, M. E.; Repka, L. M.; Reisman, S. E. J. Am. Chem. Soc. 2012, 134 (11), 5131-5137. (18) Patel, R. N. Biomolecules 2013, 3 (4), 741-777. (19) Buller, A. R.; Brinkmann-Chen, S.; Romney, D. K.; Herger, M.; Murciano-Calles, J.; Arnold, F. H. Proc. Natl. Acad. Sci. U. S. A. 2015, 112 (47), 14599-14604. (20) Herger, M.; van Roye, P.; Romney, D. K.; Brinkmann-Chen, S.; Buller, A. R.; Arnold, F. H. J. Am. Chem. Soc. 2016, 138 (27), 8388-8391. (21) Murciano-Calles, J.; Romney, D. K.; BrinkmannChen, S.; Buller, A. R.; Arnold, F. H. Angew. Chem. Int. Ed. 2016, 55 (38), 11577-11581. (22) Buller, A. R.; van Roye, P.; Murciano-Calles, J.; Arnold, F. H. Biochemistry 2016, 55 (51), 7043-7046. (23) Romney, D. K.; Murciano-Calles, J.; Wehrmuller, J. E.; Arnold, F. H. J. Am. Chem. Soc. 2017, 139 (31), 10769-10776.

Page 10 of 10

(24) Buller, A. R.; van Roye, P.; Cahn, J. K. B.; Scheele, R. A.; Herger, M.; Arnold, F. H. J. Am. Chem. Soc. 2018, 140 (23), 7256-7266. (25) Boville, C. E.; Scheele, R. A.; Koch, P.; Brinkmann-Chen, S.; Buller, A. R.; Arnold, F. H. Angew. Chem. Int. Ed. 2018, 57 (45), 14764-14768. (26) Hioki, Y.; Ogasahara, K.; Lee, S. J.; Ma, J.; Ishida, M.; Yamagata, Y.; Matsuura, Y.; Ota, M.; Ikeguchi, M.; Kuramitsu, S.; Yutani, K. Eur. J. Biochem. 2004, 271 (13), 2624-35. (27) Niks, D.; Hilario, E.; Dierkers, A.; Ngo, H.; Borchardt, D.; Neubauer, T. J.; Fan, L.; Mueller, L. J.; Dunn, M. F. Biochemistry 2013, 52 (37), 6396-6411. (28) Barends, T. R. M.; Domratcheva, T.; Kulik, V.; Blumenstein, L.; Niks, D.; Dunn, M. F.; Schlichting, I. ChemBioChem 2008, 9 (7), 1024-1028. (29) Henzler-Wildman, K.; Kern, D. Nature 2007, 450 (7172), 964-972. (30) Barducci, A.; Bonomi, M.; Parrinello, M. Wiley Interdiscip. Rev. Comput. Mol. Sci. 2011, 1 (5), 826-843. (31) Laio, A.; Gervasio, F. L. Rep. Prog. Phys. 2008, 71 (12), 22. (32) Contreras-Garcia, J.; Johnson, E. R.; Keinan, S.; Chaudret, R.; Piquemal, J. P.; Beratan, D. N.; Yang, W. J. Chem. Theory. Comput. 2011, 7 (3), 625-632. (33) Chovancova, E.; Pavelka, A.; Benes, P.; Strnad, O.; Brezovsky, J.; Kozlikova, B.; Gora, A.; Sustr, V.; Klvana, M.; Medek, P.; Biedermannova, L.; Sochor, J.; Damborsky, J. PLoS Comput. Biol. 2012, 8 (10), e1002708. (34) Maria-Solano, M. A.; Serrano-Hervas, E.; Romero-Rivera, A.; Iglesias-Fernandez, J.; Osuna, S. Chem. Commun. 2018, 54 (50), 6622-6634.


Deciphering the Allosterically Driven Conformational Ensemble in

Recommend Documents